[TERMINOLOGY] rename "changesets" to "revision bundles"

[James Blackwell]

On Mon, May 29, 2006 at 10:56:24AM -0400, Aaron Bentley wrote:
> James Blackwell wrote:
> > I'll betch you my nicest pair of shoes that a term like "bundle-revisions"
> > would get get shortened to bundle before docs got written.
> > 
> > I hit the fridge (refrigerator) for a tv (television) dinner. I may be out
> > and decide instead to get in the car (carriage, horseless carriage)/ auto
> > (automobile). I'll then buy some gas (gasoline) on my way to hit the king
> > (Burger king) for a burger (hamburger).
> 
> What's interesting to me is that we don't seem to have a
> generally-accepted short form of "computer" yet.  I was betting on
> "puter", but perhaps Strong Bad's "compy" will win the day.

How about PC? mac? box?

> > Kinda the same what that people seem to be calling shared repositories
> > just repositories. I bet that eventually gets shortened even more to repo.
> 
> Absolutely.  Already has been (see the alias list for "init-repository").
> 
> But I think you're implying that we shouldn't use names that will be
> shortened, and I don't think that's so.  Everyone knows that a tv is a
> "television set", that a fridge is a "refrigerator".

Heh. I didn't mean to imply that a long name shouldn't be used just
because it'll inevitably be shortened anyways. I meant it more in the
context of "what will it get shortened to", i.e., predicting what the
future term will become.

One of the likely shortening prospects for "revision-bundle" is "bundle".
That would be a good way to fall right into a lap of a term that JAM
doesn't care a whole lot for.

> In my dialect, "guest" and "guessed" have exactly the same pronunciation
> (both end with a flap sound), but if you point that out to someone, they
> will pronounce the words unnaturally to "prove" that they are different.
>  When they do this, they're not merely slowing their speech, they're
> making different final sounds (voiced and unvoiced stops), based on
> their knowledge of the spelling of these words.
>
> I think it's in the nature of human communication that we use ambiguous
> forms whenever that is easy, and disambiguate only when context doesn't
> resolve the ambiguities.

Sure. I'd agree with you. The human animal is inherently lazy and won't do
more than is necessary to achieve it's set of goals. 

> > My favorite would be ver/verset ("a set of 'viewable emailable
> > revisions'").  Turn a set of revisions into a verset, and apply it to get
> > a additioanal set of revisions back.  I also like revset. 
> 
> I'd prefer to use "revision" rather than "version", because that is what
> is contained.  We may want to use "version" as something different from
> "revision" in bzr, and equating them here would hinder that.  In the
> software world, "version" often implies "release version", e.g. "version
> 1.0", "version 1.9", and so I think "version" is not a good replacement
> for "revision".

Fair enough. Playing with a reverisble name doesn't work in this case,
since ver already has a connotation.

> "set" I can take or leave.  To me, "set" doesn't strongly imply that the
> revisions have been joined together.  Though the similarity to
> "changeset" might be an advantage.
>
> I don't think calling them "revsets" will prevent shortening.  By
> analogy with "changeset -> cset", they'd probably be called "rsets".
> 

Is the goal not to come up with a term that expresses the collection of
multiple revisions without using just the term revisions? I can see your
point about sets though. Though sets are collections, they can be
geographically broken. Some of the hot actresses that I'd like to fu..^W
have coffee with are in Hollywood, while others are in New York.

imho there's no need for a term as concerns usability. "I mailed you my
changes", "Send me your code", "Merge my mail" would all clearly indicate
to me the operation that was being performed. Developer discussions still
have a need of a handle. The handle, if well chosen, could be adopted by
users as well.

What do you think about revpack? 

> > I'll spare you any terms about portable containers. 
> 
> Aww.  And after you proposed "basket" for "repository"?

I neither proposed nor defended it! Fair enough that the term did pop out
of my metaphorical mouth during a conversation that "shared repository" was 
a bad term when considered against the term repository in bazaar-ng and
other revision control systems. Somebody else championed the term.

As a secret between you and I (big grin)..  I did/do have a love/hate
relationship with the term. In a world in which there are "cherrypicks",
"weaves" and "knits", terms like "baskets" and "handbaskets" does make a
certain kind of sense.

That was something I learned from Tom -- That it's ok to take a unrelated
term that provides a strong metaphor. Sometimes even that much isn't
necessary:

   I looked for a recursive acronym for Something Is Not UNIX. And I tried
   all 26 letters and discovered that none of them was a word. I decided
   to make it a contraction. That way I could have a three-letter acronym,
   for Something's Not UNIX. And I tried letters, and I came across the
   word "GNU." That was it. 
                   - RMS, Free as in Freedom.


Just my two cents.

> Given that they are directions from which "knits" or "weaves" are
> produced, I'd have thought "pattern" would be a good analogy.
> 
> Aaron

-- 
My home page:   <a href="http://jblack.linuxguru.net">James Blackwell</a>
Gnupg 06357400  F-print AAE4 8C76 58DA 5902 761D  247A 8A55 DA73 0635 7400
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 191 bytes
Desc: Digital signature
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20060529/3bcaa852/attachment.pgp